Debbie Berkley Gradient OCP Effects   Microsoft Corporation, Redmond, WA dmberk@microsoft.com

Publications:

Berkley, Deborah Milam. 1994. The OCP and Gradient Data. Studies in the Linguistic Sciences 24.2. Presented at Formal Linguistics Society of Midamerica 5. University of Illinois at Urbana-Champaign. May, 1994.

Berkley, Deborah Milam. 1994. Variability in Obligatory Contour Principle Effects. Papers from the 30th Regional Meeting of the Chicago Linguistic Society, Part 1.

Berkley, Deborah Milam. 1993. Book Notice of D. Corbin (ed.), La formation des mots: Structures et interpretations. Language 69, 619-620.

Gradient OCP Effects. (Abstract for thesis defense, Sept. 18, 2000)

This dissertation is a study of Obligatory Contour Principle effects on pairs of similar consonants found near to each other in English and Latin words. Such pairs of consonants do occur, but they are found less often than would otherwise be expected. This OCP effect (reduced occurrence of pairs of similar consonants) is strongest when the consonants are separated only by a short vowel. When more segments come
between them, the OCP effect gets weaker. In other words, the closer two consonants are to each other stringwise, the less likely it is for them to be similar to each other. Furthermore, when each consonant is in a different morpheme, or in a different syllable, the OCP effect is less than it is for consonant pairs separated by the same distance but both in the same morpheme or syllable. The OCP effect is also weaker
on consonant pairs in initial syllables than on pairs in later syllables, and on consonant pairs in unstressed syllables than on pairs in stressed syllables. I claim that the strength of this OCP effect is related to the salience of the consonants as a pair. The more salient a pair of consonants is, the less likely it is that they will be similar to each other.

I also propose that the OCP is not active on consonants in two different morphemes. Instead, I suggest that the apparent OCP effect on such consonant pairs is due to pressure from the lexicon for derived words to resemble other words. Since the lexicon is characterized by a relative absence of pairs of similar consonants, derived words will also tend to avoid such pairs.